Anders J. Lawaetz

Data fusion in metabolomic cancer diagnostics.

We have recently shown that fluorescence spectroscopy of plasma samples has promising abilities regarding early detection of colorectal cancer. In the present paper, these results were further developed by combining fluorescence with the biomarkers, CEA and TIMP-1 and traditional metabolomic measurements in the form of 1H NMR spectroscopy. The results indicate that using an extensive profile established by combining such measurements together with the biomarkers is better than using single markers.

Structure-revealing data fusion.

BackgroundAnalysis of data from multiple sources has the potential to enhance knowledge discovery by capturing underlying structures, which are, otherwise, difficult to extract. Fusing data from multiple sources has already proved useful in many applications in social network analysis, signal processing and bioinformatics. However, data fusion is challenging since data from multiple sources are often (i) heterogeneous (i.e., in the form of higher-order tensors and matrices), (ii) incomplete, and (iii) have both shared and unshared components. In order to address these challenges, in this paper, we introduce a novel unsupervised data fusion model based on joint factorization of matrices and higher-order tensors.ResultsWhile the traditional formulation of coupled matrix and tensor factorizations modeling only shared factors fails to capture the underlying structures in the presence of both shared and unshared factors, the proposed data fusion model has the potential to automatically reveal shared and unshared components through modeling constraints. Using numerical experiments, we demonstrate the effectiveness of the proposed approach in terms of identifying shared and unshared components. Furthermore, we measure a set of mixtures with known chemical composition using both LC-MS (Liquid Chromatography - Mass Spectrometry) and NMR (Nuclear Magnetic Resonance) and demonstrate that the structure-revealing data fusion model can (i) successfully capture the chemicals in the mixtures and extract the relative concentrations of the chemicals accurately, (ii) provide promising results in terms of identifying shared and unshared chemicals, and (iii) reveal the relevant patterns in LC-MS by coupling with the diffusion NMR data.ConclusionsWe have proposed a structure-revealing data fusion model that can jointly analyze heterogeneous, incomplete data sets with shared and unshared components and demonstrated its promising performance as well as potential limitations on both simulated and real data.

Structure-revealing data fusion model with applications in metabolomics

In many disciplines, data from multiple sources are acquired and jointly analyzed for enhanced knowledge discovery. For instance, in metabolomics, different analytical techniques are used to measure biological fluids in order to identify the chemicals related to certain diseases. It is widely-known that, some of these analytical methods, e.g., LC-MS (Liquid Chromatography - Mass Spectrometry) and NMR (Nuclear Magnetic Resonance) spectroscopy, provide complementary data sets and their joint analysis may enable us to capture a larger proportion of the complete metabolome belonging to a specific biological system. Fusing data from multiple sources has proved useful in many fields including bioinformatics, signal processing and social network analysis. However, identification of common (shared) and individual (unshared) structures across multiple data sets remains a major challenge in data fusion studies. With a goal of addressing this challenge, we propose a novel unsupervised data fusion model. Our contributions are two-fold: (i) We formulate a data fusion model based on joint factorization of matrices and higher-order tensors, which can automatically reveal common and individual components. (ii) We demonstrate that the proposed approach provides promising results in joint analysis of metabolomics data sets consisting of fluorescence and NMR measurements of plasma samples in terms of separation of colorectal cancer patients from controls.

Ligand binding induces a sharp decrease in hydrophobicity of folate binding protein assessed by 1-anilinonaphthalene-8-sulphonate which suppresses self-association of the hydrophobic apo-protein.

High affinity folate binding protein (FBP) regulates as a soluble protein and as a cellular receptor intracellular trafficking of folic acid, a vitamin of great importance to cell growth and division. We addressed two issues of potential importance to the biological function of FBP, a possible decrease of the surface hydrophobicity associated with the ligand-induced conformation change of FBP, and protein-inter-protein interactions involved in self-association of hydrophobic apo-FBP. The extrinsic fluorescent apolar dye 1-anilinonaphthalene-8-sulphonate (ANS) exhibited enhanced fluorescence intensity and a blueshift of emission maximum from 510-520 nm to 460-470 nm upon addition of apo-FBP indicating binding to a strongly hydrophobic environment. Neither enhancement of fluorescence nor blueshift of ANS emission maximum occurred when folate-ligated holo-FBP replaced apo-FBP. The drastic decrease in surface hydrophobicity of holo-FBP could have bearings on the biological function of FBP since changes in surface hydrophobicity have critical effects on the biological function of receptors and transport proteins. ANS interacts with exposed hydrophobic surfaces on proteins and may thereby block and prevent aggregation of proteins (chaperone-like effect). Hence, hydrophobic interactions seemed to participate in the concentration-dependent self-association of apo-FBP which was suppressed by high ANS concentrations in light scatter measurements.

Application of Rotated PCA Models to Facilitate Interpretation of Metabolite Profiles : Commercial Preparations of St. John's Wort

This paper describes the application of orthogonal rotation of models based on principal component analysis (PCA) of (1)H nuclear magnetic resonance (NMR) spectra and high-performance liquid chromatography-photo diode array detection (HPLC-PDA) profiles of natural product mixtures using extracts of antidepressive pharmaceutical preparations of St. Johns wort as an example. (1)H-NMR spectroscopy of complex mixtures is often used in metabolomic, metabonomic and metabolite profiling studies for assessment of sample composition. Interpretation of the derived chemometric models may be complicated because several sample properties often contribute to each principal component and because the influence of individual metabolites may be shared by several principal components. Furthermore, extensive signal overlap in (1)H-NMR spectra poses additional challenges to the interpretation of PCA models derived from such data. Orthogonal rotation of PCA models derived from (1)H-NMR spectra and HPLC-PDA profiles of the extracts of St. Johns wort preparations facilitate interpretation of the model. Using the varimax criterion, rotation of loadings provides simpler conditions for understanding the influence of individual metabolites on the observed clustering. Alternatively, rotation of scores simplifies the understanding of the influence of whole metabolite profiles on the clustering of individual samples.

Fast, cross cultivar determination of total carotenoids in intact carrot tissue by Raman spectroscopy and Partial Least Squares calibration

In order to speed up the breeding of orange carrots for high carotenoid content it is imperative to develop a fast and non-destructive technique. 332 roots from 86 carrot varieties grown in 2014 at the experimental farm in Høje Taastrup (DK) form the basis of this study. All roots were measured by Raman spectroscopy. The carotenoid content of the very same roots was estimated through a wet chemistry method coupled with UV-VIS at 447nm and 540nm. For the Raman spectroscopy, measurements were made on a cross section disk approximately 10cm from the root top at three different positions in the phloem. Since the top of the carrot is intact, it may still be used for growing. The final calibration model shows an uncertainty (RMSECV) of 20.5ppm, and a R(2)=0.86. It has thus proven to be well suited for prediction of carotenoids in orange carrots, and especially for ranking them according to the content.

The interrelationship between ligand binding and thermal unfolding of the folate binding protein. The role of self-association and pH.

The present study utilized a combination of DLS (dynamic light scattering) and DSC (differential scanning calorimetry) to address thermostability of high-affinity folate binding protein (FBP), a transport protein and cellular receptor for the vitamin folate. At pH7.4 (pI=7-8) ligand binding increased concentration-dependent self-association of FBP into stable multimers of holo-FBP. DSC of 3.3μM holo-FBP showed Tm (76°C) and molar enthalpy (146kcalM(-1)) values increasing to 78°C and 163kcalM(-1) at 10μM holo-FBP, while those of apo-FBP were 55°C and 105kcalM(-1). Besides ligand binding, intermolecular forces involved in concentration-dependent multimerization thus contribute to the thermostability of holo-FBP. Hence, thermal unfolding and dissociation of holo-FBP multimers occur simultaneously consistent with a gradual decrease from octameric to monomeric holo-FBP (10μM) in DLS after a step-wise rise in temperature to 78°C≈Tm. Stable holo-FBP multimers may protect naturally occurring labile folates against decomposition or bacterial utilization. DSC established an interrelationship between diminished folate binding at pH5, especially in NaCl-free buffers, and low thermostability. Positively charged apo-FBP was almost completely unfolded and aggregated at pH5 (Tm 38°C) and holo-FBP, albeit more thermostable, was labile with aggregation tendency. Addition of 0.15M NaCl increased thermostability of apo-FBP drastically, and even more so that of holo-FBP. Electrostatic forces thus seem to contribute to a diminished thermostability at low pH. Fluorescence spectroscopy after irreversible thermal unfolding of FBP revealed a weak-affinity folate binding.

The interrelationship between ligand binding and self-association of the folate binding protein. The role of detergent-tryptophan interaction.

BACKGROUND The folate binding protein (FBP) regulates homeostasis and intracellular trafficking of folic acid, a vitamin of decisive importance in cell division and growth. We analyzed whether interrelationship between ligand binding and self-association of FBP plays a significant role in the physiology of folate binding. METHODS Self-association behavior of apo- and holo-FBP was addressed through size exclusion chromatography, SDS-PAGE, mass spectrometry, surface plasmon resonance and fluorescence spectroscopy. RESULTS Especially holo-FBP exhibits concentration-dependent self-association at pH 7.4 (pI), and is more prone to associate into stable complexes than apo-FBP. Even more pronounced was the tendency to complexation between apo-FBP and holo-FBP in accord with a model predicting association between apo and holo monomers [19]. This will lead to removal of apo monomers from the reaction scheme resulting in a weak incomplete ligand binding similar to that observed at FBP concentrations <10nM. The presence of synthetic and natural detergents normalized folate binding kinetics and resulted in appearance of monomeric holo-FBP. Fluorescence spectroscopy indicated molecular interactions between detergent and tryptophan residues located in hydrophobic structures of apo-FBP which may participate in protein associations. GENERAL SIGNIFICANCE Self-association into multimers may protect binding sites, and in case of holo-FBP even folate from biological degradation. High-affinity folate binding in body secretions, typically containing 1-10nM FBP, requires the presence of natural detergents, i.e. cholesterol and phospholipids, to avoid complexation between apo- and holo-FBP.